[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[bugzilla-daemon]

[-- Attachment #1.1: Type: text/plain, Size: 7125 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98481

            Bug ID: 98481
           Summary: [Hawaii] Radeon kernel oops with vfio-pci
           Product: DRI
           Version: XOrg git
          Hardware: PowerPC
                OS: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: DRM/Radeon
          Assignee: dri-devel@lists.freedesktop.org
          Reporter: kb9vqf@pearsoncomputing.net

When DPM is enabled (dpm=1) and a Hawaii card (Radeon R9 290X) is passed to a
QEMU virtual machine via vfio, the guest kernel oopses with the following
output:

[ 1298.394440] [drm] Initialized drm 1.1.0 20060810
[ 1298.426253] [drm] radeon kernel modesetting enabled.
[ 1298.426284] checking generic (100a0000000 1d4c00) vs hw (10130000000
10000000)
[ 1298.427014] [drm] initializing kernel modesetting (HAWAII 0x1002:0x67B0
0x1002:0x0B00 0x00).
[ 1298.427030] [drm] register mmio base: 0xE0000000
[ 1298.427034] [drm] register mmio size: 262144
[ 1298.427038] [drm] doorbell mmio base: 0x40000000
[ 1298.427042] [drm] doorbell mmio size: 8388608
[ 1298.427082] [drm:radeon_device_init [radeon]] *ERROR* Unable to find PCI I/O
BAR
[ 1298.917871] [drm:radeon_atombios_init [radeon]] *ERROR* Unable to find PCI
I/O BAR; using MMIO for ATOM IIO
[ 1298.917880] ATOM BIOS: C67101
[ 1298.917971] radeon 0000:00:03.0: VRAM: 4096M 0x0000000000000000 -
0x00000000FFFFFFFF (4096M used)
[ 1298.917975] radeon 0000:00:03.0: GTT: 2048M 0x0000000100000000 -
0x000000017FFFFFFF
[ 1298.917978] [drm] Detected VRAM RAM=4096M, BAR=256M
[ 1298.917980] [drm] RAM width 512bits DDR
[ 1298.918085] [TTM] Zone  kernel: Available graphics memory: 33505984 kiB
[ 1298.918088] [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
[ 1298.918091] [TTM] Initializing pool allocator
[ 1298.918119] [drm] radeon: 4096M of VRAM memory ready
[ 1298.918122] [drm] radeon: 2048M of GTT memory ready.
[ 1298.918141] [drm] Loading hawaii Microcode
[ 1298.918151] [drm] Internal thermal controller with fan control
[ 1298.918206] Unable to handle kernel paging request for data at address
0x0000003c
[ 1298.918210] Faulting instruction address: 0xd00000000b3bff68
[ 1298.918214] Oops: Kernel access of bad area, sig: 11 [#1]
[ 1298.918216] SMP NR_CPUS=2048 NUMA pSeries
[ 1298.918220] Modules linked in: radeon(E+) ttm(E) drm_kms_helper(E) drm(E)
fuse(E) sg(E) snd_hda_codec_hdmi(E) evdev(E) snd_hda_intel(E) snd_hda_codec(E)
snd_hda_core(E) snd_hwdep(E) snd_pcm(E) snd_timer(E) snd(E) ghash_generic(E)
gf128mul(E) syscopyarea(E) sysfillrect(E) sysimgblt(E) fb_sys_fops(E)
vmx_crypto(E) i2c_algo_bit(E) soundcore(E) ip_tables(E) x_tables(E) autofs4(E)
ext4(E) crc16(E) jbd2(E) crc32c_generic(E) fscrypto(E) ecb(E) mbcache(E)
sd_mod(E) sr_mod(E) cdrom(E) hid_generic(E) usbhid(E) hid(E) ibmvscsi(E)
scsi_transport_srp(E) virtio_net(E) virtio_pci(E) virtio_ring(E) virtio(E)
xhci_pci(E) xhci_hcd(E) usbcore(E) usb_common(E) [last unloaded: drm]
[ 1298.918268] CPU: 32 PID: 2466 Comm: modprobe Tainted: G        W   E  
4.8.0-trunk-powerpc64le #1 Debian 4.8.4-1~exp1
[ 1298.918273] task: c000000fea6b9b80 task.stack: c000000fe123c000
[ 1298.918276] NIP: d00000000b3bff68 LR: d00000000ba30c88 CTR: d00000000b3bff20
[ 1298.918280] REGS: c000000fe123ede0 TRAP: 0300   Tainted: G        W   E   
(4.8.0-trunk-powerpc64le Debian 4.8.4-1~exp1)
[ 1298.918284] MSR: 800000010280b033 <SF,VEC,VSX,EE,FP,ME,IR,DR,RI,LE,TM[E]> 
CR: 84228428  XER: 20000000
[ 1298.918296] CFAR: c0000000000acea8 DAR: 000000000000003c DSISR: 40000000
SOFTE: 1
               GPR00: d00000000ba30c88 c000000fe123f060 d00000000b40af90
c000000fea554000
               GPR04: c000000fe123f140 0000000000000000 c000000fe2c6c000
c000000fff528050
               GPR08: 0000000ffea30000 c000000fee70e000 0000000000000000
d00000000ba58d80
               GPR12: d00000000b3bff20 c00000000fb92000 0000000020000000
c000000fe123fdec
               GPR16: c000000fe8f60000 c0000000008552c0 00000000000011d9
d00000000baf0000
               GPR20: d00000000baf0000 c00000000017de10 0000000000000000
0000000000000124
               GPR24: c000000fe2c6a000 c000000fe23df100 d00000000bab1640
d00000000bab1e48
               GPR28: c000000fe8f61b70 c000000fe8f61b70 0000000000000000
c000000fe123f140
[ 1298.918345] NIP [d00000000b3bff68] drm_pcie_get_speed_cap_mask+0x48/0x150
[drm]
[ 1298.918376] LR [d00000000ba30c88] ci_dpm_init+0xa8/0x13b0 [radeon]
[ 1298.918379] Call Trace:
[ 1298.918383] [c000000fe123f060] [c000000fe123f0e0] 0xc000000fe123f0e0
(unreliable)
[ 1298.918413] [c000000fe123f0e0] [d00000000ba30c88] ci_dpm_init+0xa8/0x13b0
[radeon]
[ 1298.918445] [c000000fe123f1f0] [d00000000b981a58] radeon_pm_init+0x628/0x910
[radeon]
[ 1298.918480] [c000000fe123f290] [d00000000b9e0f84] cik_init+0x324/0x790
[radeon]
[ 1298.918509] [c000000fe123f300] [d00000000b8f2c38]
radeon_device_init+0x5a8/0xc70 [radeon]
[ 1298.918539] [c000000fe123f390] [d00000000b8f6014]
radeon_driver_load_kms+0xc4/0x280 [radeon]
[ 1298.918548] [c000000fe123f410] [d00000000b3be15c]
drm_dev_register+0xfc/0x140 [drm]
[ 1298.918555] [c000000fe123f450] [d00000000b3c0798] drm_get_pci_dev+0xf8/0x220
[drm]
[ 1298.918574] [c000000fe123f4e0] [d00000000b8f0664]
radeon_pci_probe+0x134/0x1b0 [radeon]
[ 1298.918580] [c000000fe123f570] [c00000000050471c] local_pci_probe+0x6c/0x140
[ 1298.918583] [c000000fe123f600] [c0000000005055a8]
pci_device_probe+0x168/0x200
[ 1298.918588] [c000000fe123f660] [c0000000005b4900]
driver_probe_device+0x240/0x550
[ 1298.918591] [c000000fe123f6f0] [c0000000005b4d7c]
__driver_attach+0x16c/0x170
[ 1298.918595] [c000000fe123f770] [c0000000005b100c]
bus_for_each_dev+0x9c/0x110
[ 1298.918598] [c000000fe123f7c0] [c0000000005b3b5c] driver_attach+0x3c/0x60
[ 1298.918626] [c000000fe123f7f0] [c0000000005b33a8] bus_add_driver+0x308/0x390
[ 1298.918630] [c000000fe123f880] [c0000000005b5d1c] driver_register+0x9c/0x180
[ 1298.918633] [c000000fe123f8f0] [c0000000005038dc]
__pci_register_driver+0x6c/0x90
[ 1298.918640] [c000000fe123f930] [d00000000b3c0a24] drm_pci_init+0x164/0x1b0
[drm]
[ 1298.918659] [c000000fe123f9c0] [d00000000ba53440] radeon_init+0xc8/0xf8
[radeon]
[ 1298.918663] [c000000fe123fa30] [c00000000000b74c] do_one_initcall+0x6c/0x1d0
[ 1298.918667] [c000000fe123faf0] [c000000000822b50] do_init_module+0x94/0x254
[ 1298.918671] [c000000fe123fb80] [c0000000001821b0] load_module+0x2380/0x2b30
[ 1298.918674] [c000000fe123fd30] [c000000000182ca0]
SyS_finit_module+0xf0/0x170
[ 1298.918678] [c000000fe123fe30] [c000000000009560] system_call+0x38/0x108
[ 1298.918680] Instruction dump:
[ 1298.918682] f821ff81 7c7e1b78 7c9f2378 48000008 e8410018 39200000 913f0000
e93e01e8
[ 1298.918688] 2fa90000 41de0044 e9290010 ebc90038 <a13e003c> 2b891106 419e0030
2b891166
[ 1298.918696] ---[ end trace 6c565dbe73743fb5 ]---

Setting dpm=0 allows the radeon driver to load normally and the card functions
as expected under Xorg.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 8426 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[bugzilla-daemon]

[-- Attachment #1.1: Type: text/plain, Size: 290 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98481

--- Comment #1 from Timothy Pearson <kb9vqf@pearsoncomputing.net> ---
This appears related:

https://bbs.archlinux.org/viewtopic.php?pid=1537645#p1537645

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 1141 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[bugzilla-daemon]

[-- Attachment #1.1: Type: text/plain, Size: 479 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98481

--- Comment #2 from Alex Deucher <alexdeucher@gmail.com> ---
Most hypervisors do not provide access to pci config registers which the driver
needs access to to determine what pcie speeds are available.  That said, this
should work better with amdgpu and the 4.10-wip kernel:
https://cgit.freedesktop.org/~agd5f/linux/log/?h=drm-next-4.10-wip

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 1327 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[bugzilla-daemon]

[-- Attachment #1.1: Type: text/plain, Size: 520 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98481

--- Comment #3 from Timothy Pearson <kb9vqf@pearsoncomputing.net> ---
Created attachment 127594
  --> https://bugs.freedesktop.org/attachment.cgi?id=127594&action=edit
Fix radeon kernel oops when used in vfio-based guest

The amdgpu driver doesn't look quite ready for production use on SI GPUs, but I
was able to work around the issue in the radeon driver via the attached patch.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 1577 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[Bug 98481] [Hawaii] Radeon kernel oops with vfio-pci

[bugzilla-daemon]

[-- Attachment #1.1: Type: text/plain, Size: 804 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98481

Martin Peres <martin.peres@free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |MOVED
             Status|NEW                         |RESOLVED

--- Comment #4 from Martin Peres <martin.peres@free.fr> ---
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been
closed from further activity.

You can subscribe and participate further through the new bug through this link
to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/749.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 2342 bytes --]

[-- Attachment #2: Type: text/plain, Size: 159 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel